A Neural Network System for Patch Load Prediction

  • Authors:

  • Elaine T. Fonseca;Marley M. B. R. Vellasco;Marco Auré/lio C. Pacheco;Pedro C. G. Da S. Vellasco;Sebastiã/o A. L. De Andrade

  • Affiliations:

  • Civil Engineering Department, PUC-RIO, Brasil;Electrical Engineering Department, PUC-RIO, Brasil/ e-mail: marley@ele.puc-rio.br, and Computer Science and System Engineering Department, State University of Rio de Janeiro, UER ...;Electrical Engineering Department, PUC-RIO, Brasil and Computer Science and System Engineering Department, State University of Rio de Janeiro, UERJ, Brasil;Structural Engineering Department, State University of Rio de Janeiro, UERJ, Brasil/ e-mail: vellasco@uerj.br;Civil Engineering Department, PUC-RIO, Brasil, and Structural Engineering Department, State University of Rio de Janeiro, UERJ, Brasil

  • Venue:
  • Journal of Intelligent and Robotic Systems
  • Year:
  • 2001

Quantified Score

Hi-index 0.01

Visualization

Abstract

This work presents the application of Artificial Neural Networks to forecast the ultimate resistance of steel beams subjected to patch loads. A single design formula for this structural engineering problem is very difficult to obtain, due to the influence of several independent parameters. On the other hand, creating new experimental data in laboratory is very time consuming and expensive. This work demonstrates that new data can be obtained from a neural network system composed of four Back Propagation networks. The proposed neural network system presented a maximum error value lower than 15%, while the existing formulas errors were over 20%. These results confirmed the possibility of using this methodology to generate new trustworthy data. These data, coupled with experiments found in the literature, can surely help the development of a more consistent and accurate design formula.